Today for AI
HOT RADAR
llmHEAT 23.4°

AI CURATED HOT STORY · 10/23/2026

SubQ Snap 1.1 Ties for Top Decision Model Spot via Low Hallucination

1 reports archived1 independent sourcesupdated 10/23/2026, 04:02:50
Synthesis & Latest Updates
1 Sources Cross-Validated

Subquadratic released SubQ Snap 1.1, tying with d1-ultra and Elephant 9B for the top spot on the Artificial Analysis Decision Index. The model distinguishes itself with a low hallucination rate (86%) and a 384K context window, priced at $0.04 per 1M input tokens. While its raw accuracy is lower than competitors, its high non-hallucination rate drives its decision-making reliability.

LATEST/Subquadratic released SubQ Snap 1.1, tying with d1-ultra and Elephant 9B for the top spot on the Artificial Analysis Decision Index. The model distinguishes itself with a low hallucination rate (86%) and a 384K context window, priced at $0.04 per 1M input tokens. While its raw accuracy is lower than competitors, its high non-hallucination rate drives its decision-making reliability.

HEAT TRENDHourly heat curve

2 fully observed hours
Fewer than 3 fully observed hours; trend not plotted yet.

TIMELINECoverage timeline

Total 1 reports · Latest first
  1. Artificial Analysis (@ArtificialAnlys)T1·72 pts
    • SubQ Snap 1.1 scores 65 on the AA Decision Index, tying with d1-ultra and Elephant 9B, leading the next competitor by 8 points.
    • Key differentiator is low hallucination: 86% non-hallucination rate (second only to Elephant 9B), achieved by choosing 'I don't know' when uncertain.
    • Performance trade-off: Lower accuracy (20% vs 32% for d1-ultra) but significantly larger context window (384K vs 32K/16K).
    • Cost efficiency: ~$30 per 1M tasks, comparable to d1-ultra ($32) but ~3.6x more expensive than Elephant 9B ($8.4).